Predicting a supply chain performance

Abstract

Methods and systems to predict a supply chain performance are described. A system receives supply chain data for delivery of a product. The supply chain data includes input signals comprising operational plans and observed supply chain operational metrics. The input signals include a delivery date of the product. The system automatically generating predicted supply chain operational metrics across including a value at risk that is predicted for the product. The system automatically infers causal factors that impact the predicted supply chain operational metrics including impacting the value at risk that is predicted for the product. The system automatically generates action recommendations for the supply chain. An action recommendation includes a first predicted value impact and a sequence of actions impacting the product the delivery date of the product and the value at risk that is predicted for the product.

The invention claimed is: 
     
       1. A system comprising:
 one or more processors; and 
 a memory storing instructions that, when executed by the one or more processors, configure the one or more processor to perform operations comprising: 
 receiving supply chain data for delivery of a product, the supply chain data including a plurality of input signals comprising operational plans and observed supply chain operational metrics, the observed supply chain operational metrics being observed across a supply chain, the plurality of input signals including a delivery date of the product; 
 automatically generating, using a machine learning application and the operational plans and the observed supply chain operational metrics, predicted supply chain operational metrics across a plurality of nodes of the supply chain, the predicted supply chain operational metrics including a value at risk that is predicted for the product; 
 automatically inferring, using the machine learning application and the observed supply chain operational metrics, causal factors that impact the predicted supply chain operational metrics including impacting the value at risk that is predicted for the product; and 
 automatically generating, using the machine learning application based on the causal factors and the predicted supply chain operational metrics, action recommendations for the supply chain, each action recommendation having a predicted value impact, the action recommendations including a first action recommendation with a first predicted value impact, the first action recommendation further including a sequence of actions impacting the delivery date of the product and the value at risk that is predicted for the product. 
 
     
     
       2. The system of  claim 1 , wherein the receiving of the supply chain data includes automatically ingesting the supply chain data to extract the plurality of input signals. 
     
     
       3. The system of  claim 1 , wherein the predicted supply chain operational metrics include any one or more of fill rate, service level, inventory metric, demand forecast, forecast attainment, value at risk, inventory holding costs, expedite costs and liquidation costs, and wherein the predicted supply chain operational metrics are generated for at least one hierarchical level and wherein the at least one hierarchical level includes a distribution center, a hub, a logistic center, a customer location and a geographic location. 
     
     
       4. The system of  claim 1 , wherein the instructions further cause the one or more processors to perform operations comprising:
 causing presentation, within a first electronic user interface, of data representing the action recommendations and the predicted value impacts, and wherein the data representing the action recommendations and the predicted value impacts are for a product at a node in the supply chain, the action recommendations including the first action recommendation. 
 
     
     
       5. The system of  claim 4 , wherein the instructions further cause the one or more processors to perform operations comprising:
 receiving, over a network, a first selection identifying the first action recommendation from the first electronic user interface; and 
 communicating, over the network, an electronic message to a planning system, the communicating being responsive to the identifying the first action recommendation being selected, the electronic message causing an execution of a portion of a first set of actions. 
 
     
     
       6. The system of  claim 5 , wherein the instructions further cause the one or more processors to perform operations comprising:
 storing, in a tracking mechanism, the first predicted value impact, the first predicted value impact being based on the first action recommendation being executed, the storing in the tracking mechanism being in real time; and 
 regenerating, using the machine learning application and the observed supply chain operational metrics, the predicted supply chain operational metrics across the plurality of nodes of the supply chain and wherein the regenerating the predicted supply chain operational metrics includes regenerating the predicted supply chain operational metrics based on the sequence of actions impacting the product including a value at risk associated with the product at various time horizons. 
 
     
     
       7. The system of  claim 1 , wherein the machine learning application was trained with supply chain training data that was received, over time, as supply chain data from one or more supply chains including the supply chain, and wherein the supply chain data includes the operational plans and the observed supply chain operational metrics, historical actions and wherein the machine learning application utilizes a plurality of features including stock levels, stock transfer operations, supplier lead times that are stated, and supplier lead times that are actual, sales orders, stock transfer orders, and stock transfer requests. 
     
     
       8. The system of  claim 1 , wherein the instructions further cause the one or more processors to perform operations comprising:
 automatically identifying the first action recommendation as an optimal recommended action and executing the first action recommendation; 
 communicating, over a network, an electronic message to a planning system, the communicating being responsive to the identifying the first action recommendation as an optimal recommended action, the electronic message causing an execution of a portion of a first set of actions; 
 storing, in a tracking mechanism, the first predicted value impact, the first predicted value impact being based on the first action recommendation as being executed, the storing in the tracking mechanism being in real time; and 
 regenerating, using the machine learning application and the observed supply chain operational metrics, the predicted supply chain operational metrics across the plurality of nodes of the supply chain and wherein the regenerating the predicted supply chain operational metrics includes regenerating the predicted supply chain operational metrics based on the first predicted value impact. 
 
     
     
       9. The system of  claim 1 , wherein the sequence of actions impacting the value at risk of the product includes impacting the value at risk of the product at a specific time horizon and wherein the impacting the value at risk of the product at a specific time horizon includes any one or more of a first sequence of actions to expedite an outbound shipment of the product, a second sequence of actions to reallocate a quantity of outbound shipment of the product across the supply chain, a third sequence of actions to reschedule a production of the product on a first production resource, a fourth sequence of actions to reschedule a production of the product from the first production resource to a second production resource, a fifth sequence of actions to increase a quantity of the production of the product on the first production resource. 
     
     
       10. A method comprising:
 receiving supply chain data for delivery of a product, the supply chain data including a plurality of input signals comprising operational plans and observed supply chain operational metrics, the observed supply chain operational metrics being observed across a supply chain, the plurality of input signals including a delivery date of the product; 
 automatically generating, using a machine learning application and the operational plans and the observed supply chain operational metrics, predicted supply chain operational metrics across a plurality of nodes of the supply chain, the predicted supply chain operational metrics including a value at risk that is predicted for the product; 
 automatically inferring, using the machine learning application and the observed supply chain operational metrics, causal factors that impact the predicted supply chain operational metrics including impacting the value at risk that is predicted for the product; and 
 automatically generating, using the machine learning application based on the causal factors and the predicted supply chain operational metrics, action recommendations for the supply chain, each action recommendation having a predicted value impact, the action recommendations including a first action recommendation with a first predicted value impact, the first action recommendation further including a sequence of actions impacting the product the delivery date of the product and the value at risk that is predicted for the product. 
 
     
     
       11. The method of  claim 10 , wherein the receiving of the supply chain data includes automatically ingesting the supply chain data to extract the plurality of input signals. 
     
     
       12. The method of  claim 10 , wherein the predicted supply chain operational metrics include any one or more of fill rate, service level, inventory metric, demand forecast, forecast attainment, value at risk, inventory holding costs, expedite costs and liquidation costs, and wherein the predicted supply chain operational metrics are generated for at least one hierarchical level and wherein the at least one hierarchical level includes a distribution center, a hub, a logistic center, a customer location and a geographic location. 
     
     
       13. The method of  claim 10 , further comprising:
 causing presentation, within a first electronic user interface, of data representing the action recommendations and the predicted value impacts, and wherein the data representing the action recommendations and the predicted value impacts are for a product at a node in the supply chain, the action recommendations including the first action recommendation. 
 
     
     
       14. The method of  claim 13 , further comprising:
 receiving, over a network, a first selection identifying the first action recommendation from the first electronic user interface; and 
 communicating, over the network, an electronic message to a planning system, the communicating being responsive to the identifying the first action recommendation being selected, the electronic message causing an execution of a portion of a first set of actions. 
 
     
     
       15. The method of  claim 14 , further comprising:
 storing, in a tracking mechanism, the first predicted value impact, the first predicted value impact being based on the first action recommendation being executed, the storing in the tracking mechanism being in real time; and 
 regenerating, using the machine learning application and the observed supply chain operational metrics, the predicted supply chain operational metrics across the plurality of nodes of the supply chain and wherein the regenerating the predicted supply chain operational metrics includes regenerating the predicted supply chain operational metrics based on the sequence of actions impacting the product including the delivery date of the product. 
 
     
     
       16. The method of  claim 10 , wherein the machine learning application was trained with supply chain training data that was received, over time, as supply chain data from one or more supply chains including the supply chain, and wherein the supply chain data includes the operational plans and the observed supply chain operational metrics, and wherein the machine learning application utilizes a plurality of features including stock levels, supplier lead times that are stated, and supplier lead times that are actual, sales orders, stock transfer orders, and stock transfer requests. 
     
     
       17. The method of  claim 10 , further comprising:
 automatically identifying the first action recommendation as an optimal recommended action and executing the first action recommendation; 
 communicating, over a network, an electronic message to a planning system, the communicating being responsive to the identifying the first action recommendation as an optimal recommended action, the electronic message causing an execution of a portion of a first set of actions; 
 storing, in a tracking mechanism, the first predicted value impact, the first predicted value impact being based on the first action recommendation as being executed, the storing in the tracking mechanism being in real time; and 
 regenerating, using the machine learning application and the observed supply chain operational metrics, the predicted supply chain operational metrics across the plurality of nodes of the supply chain and wherein the regenerating the predicted supply chain operational metrics includes regenerating the predicted supply chain operational metrics based on the first predicted value impact. 
 
     
     
       18. The method of  claim 10 , wherein the sequence of actions impacting the delivery date of the product includes any one or more of a first sequence of actions to expedite an outbound shipment of the product, a second sequence of actions to reallocate a quantity of outbound shipment of the product, a third sequence of actions to reschedule a production of the product on a first production resource, a fourth sequence of actions to reschedule a production of the product from the first production resource to a second production resource, a fifth sequence of actions to increase a quantity of the production of the product on the first production resource. 
     
     
       19. A non-transitory machine-readable medium, the machine-readable medium including instructions that when executed by a computer, cause the computer to perform operations comprising:
 receiving supply chain data for delivery of a product, the supply chain data including a plurality of input signals comprising operational plans and observed supply chain operational metrics, the observed supply chain operational metrics being observed across a supply chain, the plurality of input signals including a delivery date of the product; 
 automatically generating, using a machine learning application and the operational plans and the observed supply chain operational metrics, predicted supply chain operational metrics across a plurality of nodes of the supply chain, the predicted supply chain operational metrics including a value at risk that is predicted for the product; 
 automatically inferring, using the machine learning application and the observed supply chain operational metrics, causal factors that impact the predicted supply chain operational metrics including impacting the value at risk that is predicted for the product; and 
 automatically generating, using the machine learning application based on the causal factors and the predicted supply chain operational metrics, action recommendations for the supply chain, each action recommendation having a predicted value impact, the action recommendations including a first action recommendation with a first predicted value impact, the first action recommendation further including a sequence of actions impacting the product the delivery date of the product and the value at risk that is predicted for the product. 
 
     
     
       20. The non-transitory machine-readable medium of  claim 19 , wherein the receiving of the supply chain data includes automatically ingesting the supply chain data to extract the plurality of input signals.